Retrieve Pulse Results (Beta)

gethttps://statsigapi.net/console/v1/experiments/{id}/pulse_results

Authorizations

STATSIG-API-KEYstringrequired

apiKey (header)

Path parameters

idstringrequired

Query parameters

controlstringrequired

Control Group ID

teststringrequired

Test Group ID

cupedstring

Whether to apply CUPED. Allowed values are "true" or "false".

confidencestring

Confidence interval (0-100)

applyBonferroniPerVariantstring

Whether to apply Bonferroni Per Variant. Allowed values are "true" or "false".

applyBonferroniPerMetricstring

Whether to apply Bonferroni Per Metric. Allowed values are "true" or "false".

bonferroniPrimaryMetricWeightstring

α allocated to primary metrics

applyBenjaminiHochbergPerMetricstring

Whether to apply Benjamini-Hochberg Correction Per Metric. Allowed values are "true" or "false".

applyBenjaminiHochbergPerVariantstring

Whether to apply Benjamini-Hochberg Correction Per Variant. Allowed values are "true" or "false".

benjaminiPrimaryMetricsOnlystring

Whether Benjamini-Hochberg correction should apply only to primary metrics. Allowed values are "true" or "false".

datestring

Date for pulse results. format must be YYYY-MM-DD

benjaminiAcrossAllTestGroupsstring

Whether to compute the Benjamini-Hochberg / Bonferroni-per-group correction across all of the experiment's test groups (matching the Console Results tab default for multi-group experiments) rather than only the requested test-vs-control pair. The returned results are always scoped to the requested pair; this controls only the correction family. Allowed values are "true" or "false". Defaults to "true".

Response

application/json

200· Get Pulse Results Success

messagestringrequired

A simple string explaining the result of the operation.

dataobjectrequired

Show child attributes

dsstringrequired

primaryMetricsobject[]required

metricIDstringrequired

metricNamestringrequired

directionalitystringrequired

Indicates the desired change direction for the metric. Use "increase" for positive changes and "decrease" for negative changes.

Allowed values:increasedecrease

Example:increase

erroroneOf

Machine-readable reason when pulse statistics are unavailable. Common values include no_data (insufficient data), echidna_unripe (not yet loaded), and setup_incomplete (misconfigured metric source). Additional string values may appear for cohort/bake windows or other load failures.

Example:no_data

Constraints:nullablestringstring

warningsstring[]

Human-readable warnings about metric data quality, load state, or partial pipeline failures.

absoluteChangenumber

confidenceIntervalDeltanumber

Absolute confidence interval delta

confidenceIntervalobject

Confidence interval bounds

Example:{"lower":-1.33,"upper":1.33}

lowernumberrequired

uppernumberrequired

percentChangenumber

sequentialTestingConfidenceIntervalDeltanumber

sequentialTestingConfidenceIntervalobject

Confidence interval bounds

Example:{"lower":-1.33,"upper":1.33}

lowernumberrequired

uppernumberrequired

percentConfidenceIntervalDeltanumber

percentConfidenceIntervalobject

Confidence interval bounds

Example:{"lower":-1.33,"upper":1.33}

lowernumberrequired

uppernumberrequired

percentSequentialTestingConfidenceIntervalDeltanumber

percentSequentialTestingConfidenceIntervalobject

Confidence interval bounds

Example:{"lower":-1.33,"upper":1.33}

lowernumberrequired

uppernumberrequired

testMeannumber

controlMeannumber

testStdnumber

controlStdnumber

testUnitsnumber

controlUnitsnumber

pValuenumber

sequentialTestingPValuenumber

Constraints:nullable

adjustedAlphanumber

reversePowernumber

percentCredibleIntervalobject

Bayesian credible interval for the percent change. Only returned when results are computed with Bayesian analysis.

Example:{"lower":-1.33,"upper":1.33}

lowernumberrequired

uppernumberrequired

expectedLossnumber

Bayesian expected loss. Only returned when results are computed with Bayesian analysis.

chanceToBeatnumber

Bayesian probability that the test beats the control. Only returned when results are computed with Bayesian analysis.

sequentialTestingLikelihoodRationumber

SPRT likelihood ratio of the observed data under the alternative vs the null. This is the same value Pulse uses.

sequentialTestingLogLikelihoodRationumber

Natural log of the SPRT likelihood ratio Pulse uses. Compare this to sprtLogLowerLimit and sprtLogUpperLimit.

sprtLowerLimitnumber

Lower SPRT decision boundary on the likelihood-ratio scale. Crossing below this accepts the null.

sprtUpperLimitnumber

Upper SPRT decision boundary on the likelihood-ratio scale. Crossing above this accepts the alternative.

sprtLogLowerLimitnumber

Lower SPRT decision boundary on the log-likelihood-ratio scale. Crossing below this accepts the null.

sprtLogUpperLimitnumber

Upper SPRT decision boundary on the log-likelihood-ratio scale. Crossing above this accepts the alternative.

sprtDecisionstring

SPRT decision from the log likelihood ratio. alternative accepts H1, null accepts H0, and continue means the ratio is still between the boundaries.

Allowed values:alternativenullcontinue

isSprtSignificantboolean

True when the SPRT decision is alternative, meaning the result is statistically significant.

sprtStatsignessstring

Direction of the SPRT decision relative to the metric's desired direction. neutral when the test has not accepted the alternative.

Allowed values:positivenegativeneutral

absoluteToplineImpactnumber

absoluteToplineImpactDeltanumber

relativeToplineImpactnumber

relativeToplineImpactDeltanumber

projectedAbsoluteToplineImpactnumber

projectedAbsoluteToplineImpactDeltanumber

projectedRelativeToplineImpactnumber

projectedRelativeToplineImpactDeltanumber

secondaryMetricsobject[]required

metricIDstringrequired

metricNamestringrequired

directionalitystringrequired

Indicates the desired change direction for the metric. Use "increase" for positive changes and "decrease" for negative changes.

Allowed values:increasedecrease

Example:increase

erroroneOf

Example:no_data

Constraints:nullablestringstring

warningsstring[]

Human-readable warnings about metric data quality, load state, or partial pipeline failures.

absoluteChangenumber

confidenceIntervalDeltanumber

Absolute confidence interval delta

confidenceIntervalobject

Confidence interval bounds

Example:{"lower":-1.33,"upper":1.33}

lowernumberrequired

uppernumberrequired

percentChangenumber

sequentialTestingConfidenceIntervalDeltanumber

sequentialTestingConfidenceIntervalobject

Confidence interval bounds

Example:{"lower":-1.33,"upper":1.33}

lowernumberrequired

uppernumberrequired

percentConfidenceIntervalDeltanumber

percentConfidenceIntervalobject

Confidence interval bounds

Example:{"lower":-1.33,"upper":1.33}

lowernumberrequired

uppernumberrequired

percentSequentialTestingConfidenceIntervalDeltanumber

percentSequentialTestingConfidenceIntervalobject

Confidence interval bounds

Example:{"lower":-1.33,"upper":1.33}

lowernumberrequired

uppernumberrequired

testMeannumber

controlMeannumber

testStdnumber

controlStdnumber

testUnitsnumber

controlUnitsnumber

pValuenumber

sequentialTestingPValuenumber

Constraints:nullable

adjustedAlphanumber

reversePowernumber

percentCredibleIntervalobject

Bayesian credible interval for the percent change. Only returned when results are computed with Bayesian analysis.

Example:{"lower":-1.33,"upper":1.33}

lowernumberrequired

uppernumberrequired

expectedLossnumber

Bayesian expected loss. Only returned when results are computed with Bayesian analysis.

chanceToBeatnumber

Bayesian probability that the test beats the control. Only returned when results are computed with Bayesian analysis.

sequentialTestingLikelihoodRationumber

SPRT likelihood ratio of the observed data under the alternative vs the null. This is the same value Pulse uses.

sequentialTestingLogLikelihoodRationumber

Natural log of the SPRT likelihood ratio Pulse uses. Compare this to sprtLogLowerLimit and sprtLogUpperLimit.

sprtLowerLimitnumber

Lower SPRT decision boundary on the likelihood-ratio scale. Crossing below this accepts the null.

sprtUpperLimitnumber

Upper SPRT decision boundary on the likelihood-ratio scale. Crossing above this accepts the alternative.

sprtLogLowerLimitnumber

Lower SPRT decision boundary on the log-likelihood-ratio scale. Crossing below this accepts the null.

sprtLogUpperLimitnumber

Upper SPRT decision boundary on the log-likelihood-ratio scale. Crossing above this accepts the alternative.

sprtDecisionstring

SPRT decision from the log likelihood ratio. alternative accepts H1, null accepts H0, and continue means the ratio is still between the boundaries.

Allowed values:alternativenullcontinue

isSprtSignificantboolean

True when the SPRT decision is alternative, meaning the result is statistically significant.

sprtStatsignessstring

Direction of the SPRT decision relative to the metric's desired direction. neutral when the test has not accepted the alternative.

Allowed values:positivenegativeneutral

absoluteToplineImpactnumber

absoluteToplineImpactDeltanumber

relativeToplineImpactnumber

relativeToplineImpactDeltanumber

projectedAbsoluteToplineImpactnumber

projectedAbsoluteToplineImpactDeltanumber

projectedRelativeToplineImpactnumber

projectedRelativeToplineImpactDeltanumber

otherMetricsobject[]

metricIDstringrequired

metricNamestringrequired

directionalitystringrequired

Indicates the desired change direction for the metric. Use "increase" for positive changes and "decrease" for negative changes.

Allowed values:increasedecrease

Example:increase

erroroneOf

Example:no_data

Constraints:nullablestringstring

warningsstring[]

Human-readable warnings about metric data quality, load state, or partial pipeline failures.

absoluteChangenumber

confidenceIntervalDeltanumber

Absolute confidence interval delta

confidenceIntervalobject

Confidence interval bounds

Example:{"lower":-1.33,"upper":1.33}

lowernumberrequired

uppernumberrequired

percentChangenumber

sequentialTestingConfidenceIntervalDeltanumber

sequentialTestingConfidenceIntervalobject

Confidence interval bounds

Example:{"lower":-1.33,"upper":1.33}

lowernumberrequired

uppernumberrequired

percentConfidenceIntervalDeltanumber

percentConfidenceIntervalobject

Confidence interval bounds

Example:{"lower":-1.33,"upper":1.33}

lowernumberrequired

uppernumberrequired

percentSequentialTestingConfidenceIntervalDeltanumber

percentSequentialTestingConfidenceIntervalobject

Confidence interval bounds

Example:{"lower":-1.33,"upper":1.33}

lowernumberrequired

uppernumberrequired

testMeannumber

controlMeannumber

testStdnumber

controlStdnumber

testUnitsnumber

controlUnitsnumber

pValuenumber

sequentialTestingPValuenumber

Constraints:nullable

adjustedAlphanumber

reversePowernumber

percentCredibleIntervalobject

Bayesian credible interval for the percent change. Only returned when results are computed with Bayesian analysis.

Example:{"lower":-1.33,"upper":1.33}

lowernumberrequired

uppernumberrequired

expectedLossnumber

Bayesian expected loss. Only returned when results are computed with Bayesian analysis.

chanceToBeatnumber

Bayesian probability that the test beats the control. Only returned when results are computed with Bayesian analysis.

sequentialTestingLikelihoodRationumber

SPRT likelihood ratio of the observed data under the alternative vs the null. This is the same value Pulse uses.

sequentialTestingLogLikelihoodRationumber

Natural log of the SPRT likelihood ratio Pulse uses. Compare this to sprtLogLowerLimit and sprtLogUpperLimit.

sprtLowerLimitnumber

Lower SPRT decision boundary on the likelihood-ratio scale. Crossing below this accepts the null.

sprtUpperLimitnumber

Upper SPRT decision boundary on the likelihood-ratio scale. Crossing above this accepts the alternative.

sprtLogLowerLimitnumber

Lower SPRT decision boundary on the log-likelihood-ratio scale. Crossing below this accepts the null.

sprtLogUpperLimitnumber

Upper SPRT decision boundary on the log-likelihood-ratio scale. Crossing above this accepts the alternative.

sprtDecisionstring

SPRT decision from the log likelihood ratio. alternative accepts H1, null accepts H0, and continue means the ratio is still between the boundaries.

Allowed values:alternativenullcontinue

isSprtSignificantboolean

True when the SPRT decision is alternative, meaning the result is statistically significant.

sprtStatsignessstring

Direction of the SPRT decision relative to the metric's desired direction. neutral when the test has not accepted the alternative.

Allowed values:positivenegativeneutral

absoluteToplineImpactnumber

absoluteToplineImpactDeltanumber

relativeToplineImpactnumber

relativeToplineImpactDeltanumber

projectedAbsoluteToplineImpactnumber

projectedAbsoluteToplineImpactDeltanumber

projectedRelativeToplineImpactnumber

projectedRelativeToplineImpactDeltanumber

Retrieve Pulse Results (Beta)

curl -X GET "https://statsigapi.net/console/v1/experiments/{id}/pulse_results" \
  -H "Content-Type: application/json" \
  -H "STATSIG-API-KEY: YOUR_API_KEY"

Was this helpful?